Modelling of particle transport, erosion and deposition in power plant gas paths

Flow evolving in turbomachinery applications is turbulent and laden by particles, such as dust, ash, dirt, etc. This would affect the behaviour of the turbine components given that solid particles can impact and then bounce off, deposit or erode their surfaces. Erosion and deposition phenomena may seriously affect the components performance, because they alter the blade profile and hence the flow field. It is thus clear that the prediction of these phenomena would be of great help form both design optimization and maintenance of turbomachinery. Besides experiments, in the last decade CFD became one of the main tool for studying flow evolution within turbine components, phenomena that involve them, and prediction of problems. In particle-laden flows, CFD is used to simulate the flow field, but also solid particle transport and dispersion, impact mechanics, adhesion or rebound, and erosion. Several approaches can be used depending on the kind of application studied and information expected. Particle transport can be simulated adopting a single or cluster of particle tracking approach (Crowe et al, 2006). Since to have a statistically independent results a large number of simulated particles is needed, the for-mer approach can be used when the domain size is not too large; moreover the instantaneous flow field is needed, otherwise turbulent dispersion of particles has to be accounted for. The cluster of particles approach (i.e., Particle Cloud Tracking model, Baxter 1989) overcomes some of these problems, since it usually uses a model for particle dispersion, computing very few trajectories to simulate a large number of particles. Particle impact/rebound and deposit/erosion are modelled using one of the available choice. For instance, impact mechanics can be modelled according to the Johnson-Kendall-Roberts theory (1971) if the particle temperature is not large enough to modify the physical properties of the particles, or the Thornton and Ning variation (1997). When the effect of temperature becomes relevant, a temperature based sticking model is used, such as that of Walsh et al. (1990). Erosion can be studied according to the model of Tabackof (1979). Aim of this study is showing how CFD can be used to simulate particle deposition/erosion in all the components of a turbine (i.e. fan, turbine, compressor), and predict the most critical regions of a given component. This will be done introducing the numerical models used for some applications, describing reference test cases, and showing/discussing results